The Agent Skills Directory

[COMMAND_EXECUTION]: The resources/implementation-playbook.md file provides Python scripts that execute management operations against cloud infrastructure APIs. For instance, the AutomatedRightsizer class contains logic to programmatically stop EC2 instances and modify their attributes using the boto3 library. Similar management capabilities are provided for S3 lifecycle policies and Lambda function concurrency settings.\n- [EXTERNAL_DOWNLOADS]: The implementation playbook includes dashboard templates that reference JavaScript libraries from well-known services. Specifically, it uses https://cdn.jsdelivr.net/npm/chart.js for rendering optimization charts and https://d3js.org/d3.v7.min.js for data visualization. These references are used for standard reporting and visualization purposes.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes data from cloud environments that could be influenced by external actors. \n
Ingestion points: The skill ingests data through cloud provider APIs, including resource tags and metadata (e.g., in AutoTagger.determine_tags and CloudCostAnalyzer._analyze_by_service), which can be controlled by any user with tagging permissions.\n
Boundary markers: There are no explicit boundary markers or instructions in the provided code samples to distinguish between trusted data and potentially untrusted embedded instructions in metadata.\n
Capability inventory: The skill possesses significant capabilities to modify production environments, including stopping/starting instances, modifying resource configurations, and changing auto-scaling/concurrency parameters.\n
Sanitization: The provided implementation logic does not include visible sanitization or validation of metadata content before it is used in decision-making processes or report generation.

database-cloud-optimization-cost-optimize